Description of Components Description of Components
Antenna: 6 Element quad scaled for 100 MHz from the ARRL Handbook. Based on a preliminary scan of locally unoccupied channels, you should design your antenna to favor the channel or group of channels you are most likely to monitor. Why a quad? For a given gain a quad may prove more convenient to place and mount. Being full-wave antennas, quads are also somewhat broader-banded than conventionally designed Yagis and are easier to match to the transmission line. Quads are less influenced in their performance by nearby objects, which simplifies locating them. All of this is not to disparage the much more common Yagi but to point out possible benefits of an alternative.
Construction is 4 inch ABS plastic sewer pipe for the boom (supported by a light wooden frame), electrical PVC conduit for cross-piece arms and 0.25 inch copper tubing for the elements. Antenna construction is where you can exercise considerable ingenuity. With all-plastic support and copper elements sprayed with Krylon, this antenna has proved impervious to weather over 6 years of operation. It has proved an exceptional performer.
The ARRL recommends trying your quad without a balun. This is good advice for radio amateurs whose objectives are unlikely to be compromised by some minor contribution of the transmission line to the antenna pattern. Most meteor work is similarly uncritical but the meticulous experimenter might want to optimize the performance of his antenna with a balun.
Here's a good trick for getting balun parts: Go to Radio Shack, buy a TV 2-way splitter and carefully pry it open (you will need to re-seal it later). Remove the insides, discarding the fine wires but keep the toroid "donut" which will have the right properties for TV/FM. Wind the balun on that and re-install with the balanced inputs connected to the splitter outputs and the unbalanced output at the splitter's single input. The case itself takes the coax shield and unbalanced ground, of course. Reseal the box.
Solder two wires to the output of the quad's driven element and run those over as short a distance as possible right into the splitter output pair. If you do this before sealing the case you can solder them right to the balun on the inside by pushing through the "F" socket. The splitter case and connectors are designed for outdoor use and are easily sealed/covered against weather. You will have to use a much-maligned "F" connector on the co-ax but if you obtain the gold-plated variety and secure it with a wrench, it will work as well as anything else. The ARRL literature describes baluns and how to design them.
FM Receiver: Harmon Kardon TU915 component FM tuner. This tuner is over 20 years old. For meteor work, the older the better is often a good idea. That's because vintage tuners incorporate discrete components making it easier to locate the point where you can draw off the Intermediate Frequency (IF) signal at 10.7 MHz. The TU915 has three highly visible IF stages marked by blue plastic-encapsulated ceramic filters. The IF signal should be taken from the output of the last filter. To this point in the circuitry an FM tuner is really no different from an AM tuner except that it admits a wider range of frequencies. The idea is to extract the signal before it reaches the limiter network. In MITROS we do our own detection and processing at that point. The audio output is necessary for audible monitoring, of course, but largely useless for analytical work.
Although we have not tried this, it should be possible to intercept the 10.7 MHz signal in a highly integrated, modern tuner by using a small coil (possibly with rod-type ferrite core) firmly fixed in position at some experimentally determined favorable spot inside the chassis. This would require considerable experimentation and some lab equipment, such as a broadband oscilloscope (a shortwave receiver tuned to 10.7 MHz and using the coil as an antenna might actually be better).
HF Receiver: For diversity reception or monitoring ionospheric activity, this is optionally available on one of the A/D channels. The receiver has been modified to buffer the S-meter signal and set a level compatible with the converter. In this way it is possible to monitor and record ionospheric activity by tuning in a station such as WWV at 15 MHz. Alternatively, a vacant frequency may be used to record and monitor thunderstorm activity or other ambient noise. For clarity, connection is omitted from the block diagram. This would simply be a line to one of the A/D channels with WinStreamMX writing a file and/or producing a display on the monitor.
Buffer: The signal at the final IF stage of the FM receiver is at a high impedance and cannot drive a transmission line of appreciable length. A simple broadband (tuning is unnecessary and only invites oscillation) amplifer using the MC1350 chip solves this problem. You can find a good circuit in "Mastering Radio Frequency Circuits" by Joseph J. Carr (McGraw-Hill). This will drive a co-ax to several meters with absolute stability.
10.7 MHz Amplifier: The receiver IF output needs a big boost to become useful. This module supplies approximately 65 dB of low noise gain with excellent dynamic range. The circuit is proprietary and was kindly donated to the project by Interceptor Electronics (Tom Dawson, WB3AKD) of Round Hill, Va. You can find one based on similar principles in "Mastering Radio Frequency Circuits" and it should perform almost as well. The inputs and outputs are untuned but a 10.7 MHz FM filter is required between the two amplifying stages. You can obtain an excellent filter from Mini-Circuits for a few dollars. It is important to accommodate the full bandwidth of an FM channel to avoid program modulation of the received signal power.
This is a good place to explain how/why FM transmissions are used in the first place ... Amplitude Modulated (AM) signals such as TV video transmissions (the sound is FM) present a varying amount of power to the receiving station, dependent upon program content. The power in an FM channel is largely independent of modulation characteristics because the information is represented by the frequency deviation of the carrier signal. That means if you take care to intercept the full bandwidth of an FM transmission, you obtain a signal whose strength is largely dependent on the properties of the medium it passes through or is scattered by (like a meteor generated plasma column). This facilitates critical examination of scattered power.
An important drawback is that it is very difficult (no one, to our knowledge, has demonstrated how to do it) to utilize an FM transmission for meteor velocity determination. Injecting a fixed-offset "Beat Frequency" source into the IF channel on behalf of this analysis would seem meaningless. The carrier of an AM transmission remains fixed in frequency, subject to Doppler effects, providing both audible and analytical cues to the velocity vector component aligned with the receiving path.
Analog Multiplier: This circuit, based on the MC1496 chip, was adapted from one distributed by Chuck Forster to members of the Society of Amateur Radio Astronomers. MITROS changes include a network for nulling the output offset. The circuit achieves detection by squaring all the input values. The output is, therefore, proportional to power and capable of enormous dynamic range compared to a diode. Another advantage over diodes is that temperature compensation is unnecessary in most situations for accurate performance.
As an alternative approach to a multiplier we have experimented with temperature compensated diode circuits and found them very good. In addition to temperature compensation, the diode must be statically biased into forward conduction or it will be insensitive to the weakest signals. Complexity of construction is similar to a multiplier circuit. The chief drawback to diode based circuits is limited dynamic range.
Variable Integrator: Output from the detector comprises the full IF bandwidth plus components resulting from the multiplier action. This is unsuitable for display or analytical purposes. The integrator continuously sums detector output over user selectable time intervals, settable by a front panel switch.
Elliptical Filter: The 9325 chip provides the foundation for this circuit which is adapted from the product application notes. The filter provides a very steep 60 dB/octave cut-off at front panel selectable frequencies of 40 Hz, 100 Hz and at 100 Hz steps thereafter up to 1000 Hz. With proper selection, this filter minimizes the possibility of aliasing by the analog to digital converter which follows. The MITROS version of this circuit employs "cold switching" with analog CMOS switches to minimize introduction of noise at this stage. The filter output is essentially the RF envelope, bandwidth limited to satisfy sampling requirements. This is the primary MITROS output and is continuously displayed on the computer monitor.
The filter is available only for the RF envelope channel. The ATRG and Counter outputs are already bandwidth limited by their own circuitry.
Despite the shortcut afforded by a specialized integrated circuit, building an elliptical filter with selectable cutoffs is a tedious, unpleasant exercise, owing to the large number of discrete components still required. The circuit must be clocked as well, requiring a video color burst crystal and supporting circuitry, including anti-aliasing network. The performance is exceptional, however.
Auto-Tracking Reference Generator (ATRG): The ATRG, in combination with the Event Detector provides "smart" detection of meteor events using only hardware. The ATRG continuously develops a voltage dependent on background noise at a variable rate of computation, offset and averaging interval controlled by the equipment operator using front-panel settings (once adjusted, these seldom require additional attention). This voltage is fed back to the Event Detector which uses it as a threshold for determining when a meteor event is in progress. Upon detection of an event, the Event Detector raises a TTL level which goes back to the ATRG. This signal is used to prevent updating the threshold computation until the event has ended. In this way meteor events minimally influence the detection threshold.
Should an event last longer than some pre-determined interval of time (operator selectable but typically a couple of minutes), the ATRG assumes this is not a valid event at all but proof of a persistent change in the noise level. It then automatically advances the threshold to the maximum possible value (to ensure it is above the new noise level) and initiates a RESET operation which lets the threshold fall until originally established averaging interval and offset criteria are met whereupon automatic tracking resumes.
In this way the ATRG responds not only to gradual fluctuations in background level, but also abrupt shifts such as local RFI, more stations coming on-air, etc. The count is lost, of course, during the reset operation but this lasts only for a few seconds. The ATRG is analogous to AGC in a receiver, but differs with its logic feedback control and ability to suspend operation when something important must be passed through without modification. The ATRG operates on the integrated signal and is implemented using LF353 operational amplifiers and low-power TTL. The ATRG makes unattended operation of MITROS possible.
This circuit was modeled on actual human performance with manual controls when attempting to set thresholds in presence of varying noise levels.
The ATRG reference function can be displayed on the computer monitor and/or written to a file.
** More recently, the ATRG function has been duplicated in software and in some respects improved upon.
Event Detector: This module continuously compares the integrator output to a reference level (almost always from the ATRG although an operator adjustable static internal level is available). An amplitude hysteresis control may be set to stabilize the circuit against nuisance triggering in presence of noise that gets past the integrator. It also provides a degree of "stickiness" to help avoid losing an event prematurely.
When the reference is exceeded, a precision ramp generator is triggered. If the "event" expires before the ramp reaches a pre-determined level then it is assumed not to have been an event at all and is ignored - the ramp generator is reset. Otherwise, a valid event is determined to be in progress and an output TTL line is asserted. This line is asserted for the duration of the event and is the same one used to clamp the ATRG reference computation. It also drives the Periodic Counter.
Audio Gate: The Event Detector incorporates an audio gate. This is simply an analog switch to which any standard audio source may be connected. This is usually the audio output from the FM receiver but it could be some other signal, like a continuous tone, to indicate an event is in progress. When the event signal is asserted, the switch closes and the audio is passed through to a small amplifier and loudspeaker or headphones. If the equipment operator desires monitoring the audio this circuit provides total relief from the noise that would otherwise have to be endured between events. A front-panel LED lights whenever an event is in progress, providing visual indication of event occurrences.
Periodic Counter: Every time an event TTL signal is received a precise amount of fixed charge is added to a sample-and-hold circuit. Every 5 minutes the accumulated charge is dumped. The output represents, therefore, the event count over a 5 minute period and is optionally written to a file. Software can later be used to develop a count over any other time period. This output can also drive a strip-chart recorder, enabling a display of event count even if there is no computer available.
A/D (Analog to Digital Converter): These are Universal Lab Interfaces (ULI) available from Vernier Software and Technology (superseded by another product in the last two years although the ULI is still offered). The ULI attaches to the computer by means of any serial port. The data logging software supplied by Vernier is unsuitable for many astronomical purposes because it is curiously limited to acquiring only 32000 samples. For meteor work you will often need to collect millions of samples in a single session. Accordingly, we wrote WinStreamMX which supports key ULI functions with unlimited data acquisition capability. WinStreamMX additionally provides comprehensive file headers for properly identifying an observing session and a vastly more flexible display.
Computer/Software: This is the ubiquitous "Wintel" platform. We have found that Windows XP and/or Windows 98/SE combined with the Delphi programming product (Borland) well suited to the development of robust, high-performance engineering/scientific software. The resulting application code runs without any problems and is easily maintained. Why not the more popular C++? We find Object Pascal superior in almost every respect, from support for the Object Oriented design paradigm to readability of the source code and productivity. This may not be everyone's experience, but it is ours.
WinStreamMX performs data acquisition only. A comprehensive setup panel gives the operator complete control over sampling rates (including access to the ULI hardware registers), target file locations and names, channel selection and file identifier information. A scheduler supports having a session begin automatically at some future time. You will find a screen-shot of the WinStreamMX control panel in the Image Gallery.
WinStreamSP provides signal post-acquisition processing. Anything is possible here which means the program is in a continuous state of development. Among other things - and possibilities - it supports file conversions, frequency analysis, counting and subset evaluation of meteor data.